At 9:56 PM +0200 6/19/07, Bjoern Hoehrmann wrote:
Could you point out where the draft requires making changes?
First, let me point out that the contrapositive is covered in Section 10: Clients MUST NOT assume that an Atom Entry returned in the Feed is a full representation of an Entry Resource and SHOULD perform a GET on the URI of the Member Entry before editing it.
I so far assumed you can make a conforming implementation that does little but passing through the content I want to publish as-is, your text seems to rule that out.
It is ruled out if publishing a feed would create an invalid feed, such as one with non-unique atom:id values. That is, the server is responsible for following the MUST rules of the Atom format when it is publishing, even if the client gives it something that would not follow those MUSTs.
